Question: There is a TV avid person, who wants to spend his maximum time on TV. There are N channels that telecast programs of different length at different timings. WAP to find the program and channel number so that the person can spend his max time on TV.
Condition: If he watches a program, he watches it completely.
Example:
Channel1:
prg1: 8:00 am - 9:00am
prg2: 9:30 am - 12:00 am
Channel2:
prg1: 8:15 am - 9:30am
prg2: 9:30 am - 11:45 am
In this case, between 8:00 AM to 12:00 noon, the answer is:
ch2/prg1 + ch1/prg2
Answer: This is a clear case of dynamic programming. If you remember problem of longest increasing subsequence, you will be able to see a similarity between this and LIS problem. Lets calculate that if I includes program x in my schedule then how can I spent most of my time on tv till the end of program x. So when I calculate the same for the last program among all the channels, I will be easily able to tell that what will be maximum time, I can spend on tv.
Algorithm: I have a data structure, which represents a program as below:
Struct program{ int id; //unique id to identify a program int cnt; //This will tell me what max_time, I can spend on //tv till end of this program, if I include this program. int start; // Start time of the program int end; // End time of the program int channel_no; // Channel id of the program, skipped in main code }
Embedded Document below is showing the algorithm with an Example. I didn't write algo here for the sake of indentation and example to save the length of this post.
Code:
In this code, I just worked on a merged array for all the programs and sort it. But need to do that separately in nlogk steps.
#include <iostream> using namespace std; class prog { public: int id; //unique id to identify a program int cnt; //This will tell me what max_time, I can spend on //tv till end of this program, if I include this program. int start; // Start time of the program int end; // End time of the program int channel_no; // Channel id of the program prog(char i, int count, int st, int en) { id = i; cnt = count; start = st; end = en; } }; int comp(const void *a,const void *b) { if ((*(prog *)a).end == (*(prog *)b).end) return 0; else if ((*(prog *)a).end < (*(prog *)b).end) return -1; else return 1; } /* Test Data Program id P1 P2 P3 Start time 8:00 9:00 10:30 End time 8:30 10:00 11:30 Channel 2: Program id P4 P5 P6 Start time 8:15 9:30 10:45 End time 9:15 10:15 11:30 */ int main(int argc, char* argv[]) { /* say it's the final unsorted array after merging. I will sort it using qsort. We should do a sorted merge from individual k channel arrays that will take O(nlogk) time where n is total number of programs.*/ prog programs[6]={ prog(1,0,800,850), prog(2,0,900,1000), prog(3,0,1050,1150), prog(4,0,825,925), prog(5,0,950,1025), prog(6,0,1075,1150) }; int i, max, j, n; // Printing Just for Debuging for (i=0;i<6;i++) cout << programs[i].id << " "; /*Sort the array. We should do a sorted merge from individual k channel-arrays that will take O(nlogk) time where n is total number of programs. */ qsort(programs, 6, sizeof(prog), comp) ; // Printing Just for Debuging for (i=0;i<6;i++) cout << programs[i].id << " "; n = sizeof(programs)/sizeof(prog); for (i=0; i< n; i++) { max = 0; j = 0; while (programs[i].start > programs[j].end ) { if (max < programs[j].cnt) { max = programs[j].cnt; } j++; } programs[i].cnt = max + programs[i].end - programs[i].start; } //Get the maximum count value int res = 0; for (i=0; i< n; i++) { if (res < programs[i].cnt) res = programs[i].cnt; } cout << "result =" << res << endl; return 0; }
Efficiency:
This has a few steps:
- Merge sort individual k channel arrays: O(nlogk) where n is total number of programs.
- Find count for each program. For finding count for each program, in worst case we need to traverse all the programs. So O(n). For n programs total complexity: O(n^2).
- Find maximum cnt in programs array: O(n)
So total complexity will be O(nlogk) + O(n^2) + O(n) = O(n^2)
Update: Another version of this problem can be to find maximum number of programs that person can watch.
You can also see which program he will watch by saving predecessor in another array as we do in LIS.
Subscribe - To get an automatic feed of all future posts subscribe here, or to receive them via email go here and enter your email address in the box. You can also like us on facebook and follow me on Twitter @akashag1001.
nice algo!
...Just wondering if you are going to utilize this solution in your real life over the weekends... :)
It's noteworthy that LIS has an O(n * lg n) solution which is pretty involved.
http://en.wikipedia.org/wiki/Longest_increasing_subsequence